Published on: June 11, 2026 · 8 min read
Three-to-one is the number everyone quotes and almost no one computes correctly. "We're at 3:1 LTV to CAC" gets nodded through in board decks, pitch decks, and growth reviews as if it were a law of physics. It isn't. It's a ratio of two numbers that are each easy to inflate, and the inflation always runs in the flattering direction.
The appeal is obvious. LTV:CAC compresses an entire business model into one figure: for every rupee you spend acquiring a customer, how many rupees of lifetime value come back? Above 3, the lore says you're healthy and should spend more. Below 1, you're lighting money on fire. Clean, intuitive, board-friendly.
And mostly wrong, because three different errors hide inside that ratio — and unlike payback, where the traps are in the denominator, here they compound.
It's worth saying plainly: 3:1 is a heuristic from early-2010s US SaaS, where gross margins were 80%+, sales cycles were predictable, and capital was cheap. It was never a universal constant. Drop it onto a payments business at 30% margin, or a consumer app with brutal early churn, and the same ratio means something completely different. Borrowing the benchmark without borrowing the context is the first mistake. The next three are arithmetic.
The most common way to compute LTV is from your existing customer base — average revenue, average lifespan, of the people who are still here. But the customers still here are the ones who didn't churn. You're measuring the survivors and projecting their loyalty onto everyone, including the third who left in the first two months and contributed almost nothing.
Real LTV has to be computed on a full cohort from acquisition — survivors and casualties together — or it systematically overstates. The healthier your survivors look, the more the dead ones are dragging the true number down, unseen.
LTV is, by definition, future money. Margin you'll collect in year two and three, if the customer stays and if the business is around to collect it. Future money is worth less than money in hand — that's not pessimism, it's the time value of money, and ignoring it is the same error as valuing a bond at the sum of its coupons.
You have to discount future contribution back to present value. The further out the cash flow and the riskier the business, the more aggressive the discount. Undiscounted LTV flatters long-lived-but-thin customers and makes a 5-year payback look the same as a 5-month one.
A company-wide LTV:CAC averages your best channel and your worst into one number. Referral customers might be 8:1; a particular paid channel might be 0.7:1 and actively destroying value. Blend them and you get a comfortable 3:1 that tells you nothing about where to spend the next rupee — and quietly green-lights the channel that's bleeding.
The ratio is only decision-useful per channel and per cohort. The aggregate is a vanity number.
Same customers, the naive ratio versus the honest one:
| Step | Naive | Corrected |
|---|---|---|
| Lifetime value basis | survivors only | full cohort |
| Gross LTV | ₹12,000 | ₹7,400 |
| Discounted to present value | — | ₹5,600 |
| CAC basis | blended ₹1,400 | paid ₹2,800 |
| LTV : CAC | 8.6 : 1 | 2.0 : 1 |
Illustrative, but the direction is real: every correction pulls the ratio down, and the gap between "8.6:1, spend everything" and "2.0:1, be careful" is the difference between a good quarter and a bad year.
A ratio that can only move in the flattering direction when you cut corners isn't a health metric. It's a comfort metric.
I don't throw LTV:CAC out — I make it earn its place:
Computed that way, a 2:1 can be genuinely excellent and a reported 6:1 can be a mirage. The number matters far less than the assumptions you were willing to write down next to it.
Discount rates are a judgement call, and reasonable people pick different ones. Full-cohort LTV needs cohorts old enough to have a real tail, which young companies simply don't have — so you're extrapolating, and extrapolation is where optimism creeps back in. And LTV:CAC, however carefully built, still ignores timing: a 3:1 that takes four years to realise is worse for a cash-constrained business than a 2:1 that pays back in six months. That's why I treat it as one panel of the dashboard, never the dashboard.